Gliomas are the most common primary tumors arising in the human brain. Despite surgery, chemotherapy, and radiotherapy, the most malignant glioma, glioblastoma, is almost always fatal with a median survival of less than a year and a 5-year survival of 5.5% or less. We are evaluating a novel means of possible therapy for gliomas - the use of genetically engineered viruses to specifically destroy glioma cells while leaving defective retroviral vector demonstrated the feasibility of this approach. We extended these studies to a replication-competent thymidine kinase-negative Herpes Simplex Virus 1 (HSV1) mutant and recently we have confirmed the validity of this approach with other effective in the presence of a competent immune system; 2, this can be extended to other nervous system tumors including medulloblastoma and malignant meningioma; 3, these effects can be duplicated with HSV mutants which retain thymidine kinase proficiency; and 4, with appropriate viral engineering, encephalitis can be eliminated. We now propose studies: 1, to further attenuate replication-competent herpes viruses by engineering combinations of mutations known to decrease neuropathogenicity, yet expected to retain the ability to kill brain tumor cells in vitro and in vivo; 2, to study the possible potentiation of tumor cell killing using the "by-stander" effect by adding defective HSV1 vectors containing highly expressed HSV-TK or CMB UL97 sequences and subsequent treatment with ganciclovir; 3, to construct and test host-range HSV mutants containing cell-specific promoters to kill specifically defined cell populations; 4, to determine the effect of prior exposure to HSV on the ability to HSV vectors to kill tumors; 5, to study the acute and long- term neuropathology and the dynamics of viral spread in tumor following inoculation of the most efficacious of these mutants (or mutants plus defective vector combinations) in immunocompetent mice and in primates following intracranial inculcation. We predict that efficacious yet safe replication-competent viruses can be designed to kill malignant brain tumor cells in situ without harm to surrounding brain. Active viral replication in situ will allow better tumor penetration by the virus thus overcoming some of the current limitations of using only replication- defective vector systems for gene therapy of cancer.